MiRobotEnv.py

import gymnasium as gym
from gymnasium import spaces
import numpy as np

# Constantes pour les états et les récompenses
CMD_AVANCER = 0
CMD_TOURNER = 1
ETAT_FAIM = 2
ETAT_SOMMEIL = 3
ETAT_HUMEUR = 4

# Actions
ACTION_ARRETER = 0
ACTION_AVANCER = 1
ACTION_TOURNE_G = 2
ACTION_TOURNE_D = 3


class MiRobotEnv(gym.Env):
    """Environnement de simulation pour MiRobot, un chiot robot apprenant
à réagir aux commandes du maître et à son état interne.
    """
    metadata = {"render_modes": ["human"], "render_fps": 30}

    def __init__(self):
        super(MiRobotEnv, self).__init__()

        self.action_space = spaces.Discrete(4)

        low = np.array([0.0, 0.0, 0.0, 0.0, -1.0], dtype=np.float32)
        high = np.array([1.0, 1.0, 1.0, 1.0, 1.0], dtype=np.float32)
        self.observation_space = spaces.Box(low=low, high=high, dtype=np.float32)

        self.state = None

    def reset(self, seed=None, options=None):
        super().reset(seed=seed)

        initial_state = np.array([
            0.0,
            0.0,
            self.np_random.uniform(low=0.0, high=0.2),
            self.np_random.uniform(low=0.0, high=0.2),
            self.np_random.uniform(low=-0.1, high=0.1)
        ], dtype=np.float32)

        self.state = initial_state

        info = {}
        return self.state, info

    def _update_internal_states(self):
        self.state[ETAT_FAIM] = np.clip(self.state[ETAT_FAIM] + 0.005, 0.0, 1.0)
        self.state[ETAT_SOMMEIL] = np.clip(self.state[ETAT_SOMMEIL] + 0.003, 0.0, 1.0)

        humeur_decay = -0.005 if (self.state[ETAT_FAIM] > 0.5 or self.state[ETAT_SOMMEIL] > 0.5) else 0.001
        self.state[ETAT_HUMEUR] = np.clip(self.state[ETAT_HUMEUR] + humeur_decay, -1.0, 1.0)

    def _calculate_reward(self, action):
        reward = 0.0

        if self.state[CMD_AVANCER] > 0.5:
            if action == ACTION_AVANCER:
                reward += 1.0
            elif action == ACTION_ARRETER:
                reward -= 0.5

        if self.state[CMD_TOURNER] > 0.5:
            if action == ACTION_TOURNE_G or action == ACTION_TOURNE_D:
                reward += 1.0
            elif action == ACTION_AVANCER:
                reward -= 0.5

        reward -= self.state[ETAT_FAIM] * 0.1
        reward -= self.state[ETAT_SOMMEIL] * 0.1

        reward += self.state[ETAT_HUMEUR] * 0.1

        return reward

    def _simulate_user_command(self):
        if self.np_random.random() < 0.2:
            self.state[CMD_AVANCER] = 0.0
            self.state[CMD_TOURNER] = 0.0

            choice = self.np_random.integers(0, 3)

            if choice == 1:
                self.state[CMD_AVANCER] = 1.0
            elif choice == 2:
                self.state[CMD_TOURNER] = 1.0

    def step(self, action):
        self._update_internal_states()
        reward = self._calculate_reward(action)
        self._simulate_user_command()

        terminated = False
        truncated = False
        info = {}

        return self.state, reward, terminated, truncated, info

    def render(self, mode='human'):
        faim = self.state[ETAT_FAIM] * 100
        sommeil = self.state[ETAT_SOMMEIL] * 100
        humeur = self.state[ETAT_HUMEUR]
        cmd_a = "OUI" if self.state[CMD_AVANCER] > 0.5 else "NON"
        cmd_t = "OUI" if self.state[CMD_TOURNER] > 0.5 else "NON"

        print(f"--- État MiRobot ---")
        print(f"  > Commande Avancer: {cmd_a}, Commande Tourner: {cmd_t}")
        print(f"  > Faim: {faim:.0f}%, Sommeil: {sommeil:.0f}%")
        print(f"  > Humeur: {humeur:.2f} (entre -1.0 et 1.0)")

    def close(self):
        pass